Artificial Intelligence-Based Rapid Design of Grease with Chemically Functionalized Graphene and Carbon Nanotubes as Lubrication Additives

Abstract

Rapid chemical functionalization of additives and efficient determination of their optimum concentrations are important for designing high-performance lubricants, especially under multi-additive conditions. Herein, chemically functionalized graphene (FGR) and carbon nanotubes (FCNTs) were rapidly prepared by microwave-assisted ball milling and subsequently introduced into grease as additives. The tribological properties of the additives in grease at different concentrations and ratios were measured using a four-ball test. A reliable artificial neural network (ANN) model was established according to a few test results. Subsequently, the optimal concentration of multiple additives in the grease was predicted using a genetic algorithm and experimentally validated. The results indicated that the introduction of FGR (0.14 wt %) and FCNT (0.16 wt %) improved the antifriction and anti-wear performance of the base grease by 25.66 and 29.34%, respectively. The results of the ANN model analysis and friction interface characterization indicate that such performance is principally attributed to the synergistic lubrication of the FGR and FCNT.

Bridging Nanomanufacturing and Artificial Intelligence—A Comprehensive Review

“…The performance of the grease was continuously monitored using an artificial neural network (ANN) [53] Nanomedicine where nanometric structures provide a highly sophisticated, targeted approach to diagnosing health conditions.…”

Using artificial intelligence to predict the tribology behavior of MoS2-Al2O3 hybrid nanofluid

“…ANN algorithm is inspired by natural brains' architecture and involves a number of simple but highly interconnected information processing neurons. ML algorithm can solve nonlinear and complex relationships through data training [17]. They can infer previously unknown relationships, allowing for a simplified model and estimation of hidden data.…”

“…performance of MoS 2 -Al 2 O 3 hybrid nanofluid, including two artificial neural network algorithm (multilayer perceptron (MLP) and back propagation (BP)), as well as two machine learning algorithms (random forest (RF) and k-nearest neighbors (KNN)). These four algorithms were selected with reference to relevant studies[16][17][18][19][20]. The date size, algorithm stability, actual number of variables, comprehensiveness and comparison of algorithm types also played a vital role in the selection of algorithms.…”